Detection of a New Heparin-Dependent Inhibitor of Thrombin in Human Plasma

Home , Macroglobulin, Ovalbumin, Protein C inhibitor, Trypsin inhibitor

Detection of a new heparin-dependent inhibitor of thrombin in human plasma.

D M Tollefsen, M K Blank

J Clin Invest. 1981;68(3):589-596. https://doi.org/10.1172/JCI110292.

Research Article

We have demonstrated that human plasma contains a heparin-dependent inhibitor of thrombin that is distinguishable from antithrombin III (AT III). When a 1:50 dilution of plasma was incubated with greater than or equal to 0.01 U/ml heparin and 1 U/ml 125I-thrombin, the labeled thrombin B-chains became incorporated into two complexes of Mr-96,000 and Mr- 85,000 that were separated by polyacrylamide gel electrophoresis in the presence of sodium dodecyl sulfate and beta- mercaptoethanol. Neither complex was detectable at heparin concentrations less than 0.01 U/ml. When a limiting amount of 125I-thrombin was present, the proportion of radioactivity incorporated into each of the two complexes varied with the heparin concentration. Thus, the Mr-85,000 complex predominated at 0.01-5 U/ml heparin, whereas the Mr-96,000 complex predominated at 5-100 U/ml heparin. The Mr-85,000 complex reacted with antibodies to human AT III and comigrated with the purified thrombin-AT III complex. The Mr-96,000 complex did not react with antibodies to AT III or to alpha 1-antitrypsin, and it was detected in normal quantities after incubating 125I-thrombin with plasma immunodepleted of AT III, alpha 2-antiplasmin, alpha 2-macroglobulin, C1 inactivator, alpha 1-antichymotrypsin, or inter-alpha-trypsin inhibitor. The protein that combines with thrombin to form the Mr-96,000 complex was estimated to be present at a minimum concentration of 90 +/- 26 micrograms/ml (mean +/- SD) in identical to any of the known plasma protease […]

Find the latest version: https://jci.me/110292/pdf Detection of a New Heparin-dependent Inhibitor of Thrombin in Human Plasma

DOUGLAS M. TOLLEFSEN and MARY K. BLANK, Division of Hematology-Oncology, Department of Internal Medicine, Washington University School of Medicine, St. Louis, Missouri 63110

A B S T R A C T We have demonstrated that human INTRODUCTION plasma contains a heparin-dependent inhibitor of Heparin has been known for many years to be a potent thrombin that is distinguishable from antithrombin anticoagulant and is widely used in the treatment of III (AT III). When a 1:50 dilution of plasma was incu- thromboembolic disorders in man (1). It is generally bated with .0.01 U/ml heparin and 1 U/ml 1251-throm- accepted that heparin exerts its anticoagulant effect bin, the labeled thrombin B-chains became incor- by activating the plasma protein antithrombin III porated into two complexes of Mr-96,000 and M,-85,000 (AT III)' (formerly called "heparin cofactor") (2). AT that were separated by polyacrylamide gel electro- III inhibits thrombin and other coagulation proteases phoresis in the presence of sodium dodecyl sulfate (Factors Xa, IXa, XIa, XIIa, and kallikrein) by form- and f3-mercaptoethanol. Neither complex was de- ing 1:1 stoichiometric complexes with each protease tectable at heparin concentrations <0.01 U/ml. When (2). That the complexes are stable in denaturing a limiting amount of 1251-thrombin was present, the into each of agents suggests that a covalent bond is formed be- proportion of radioactivity incorporated tween the protease and the inhibitor and allows de- the two complexes varied with the heparin concen- tection of the complexes in sodium dodecyl sulfate tration. Thus, the Mr-85,000 complex predominated (SDS) polyacrylamide gels. at 0.01-5 U/ml heparin, whereas the Mr96,000 com- It is clear that heparin accelerates the rate at which plex predominated at 5-100 U/ml heparin. The Mr, AT III reacts with various proteases in vitro. In the 85,000 complex reacted with antibodies to human case of thrombin, the second-order rate constant of com- AT III and comigrated with the purified thrombin- plex formation increases from 4.3 x 105 M-1 min-' in AT III complex. The Mr-96,000 complex did not the absence of heparin to 8 x 108 M-l min-' at optimal react with antibodies to AT III or to al-antitrypsin, heparin concentrations (3). None of the other recog- and it was detected in normal quantities after incubat- nized protease inhibitors in plasma2 is known to be ing 1251-thrombin with plasma immunodepleted of AT stimulated by heparin, with the possible exception of III, a2-antiplasmin, a2-macroglobulin, Cl inactivator, C1 inactivator (5), which does not inhibit thrombin (6). al-antichymotrypsin, or inter-a-trypsin inhibitor. The On the other hand, several previous reports have protein that combines with thrombin to form the M,- claimed to separate from AT III a thrombin inhibitor 96,000 complex was estimated to be present at a mini- with heparin cofactor activity, although the material re- mum concentration of 90+26 ,ug/ml (mean±SD) in sponsible for this activity was neither purified nor normal plasma. We conclude that the protein is not characterized by immunological techniques (7-9). identical to any of the known plasma protease in- Thus, the existence of a heparin cofactor apart from hibitors and that at relatively high heparin concen- AT III has remained in doubt. trations in vitro it reacts with thrombin more rapidly In the present investigation, we have incubated than does AT III. 125I-labeled thrombin with plasma and separated the

This work was presented in part at the Annual Meeting of the American Society for Clinical Investigation, 25-27 'Abbreviations used in this paper: AT III, antithrombin April 1981, San Francisco, and has appeared in abstract form. III; SDS, sodium dodecyl sulfate. 1981. Clin. Res. 29: 523A. 2 These include al-antitrypsin, al-antichymotrypsin, inter- Received for publication 9 March 1981 and in revised a-trypsin inhibitor, a2-antiplasmin, AT III, Cl inactivator, and form 21 April 1981. a2-macroglobulin (4).

J. Clin. Invest. (© The American Society for Clinical Investigation, Inc. * 0021-9738/81/09/0589/08 $1.00 589 Volume 68 September 1981 589-596 anitichymiiotrypsin, a2-macroglobulin, Cl inactivator, and Top of inter-a-trypsin inhibitor were purchased from Calbiochem- running gel Behring Corp., American Hoechst, La Jolla, Calif. Rabbit antiseruim directedl against human a2-antiplasmin was purchased from Nordic Immunological Laboratories (USA), Sani Clemiienite, Calif. Apiezon A oil was obtained from Mr 96,000--. James B. Biddle Co., Plymouth Meeting, Pa. Mr 85000-' Proteins. Humlan prothrombin was prepared by the method Mr 72,000-- 4 of Miletich et a]. (11). Prothrombin was activated using Taipani snake venom (12) and thrombin was subsequenltly isolate(l by chromatography on Amberlite CG-50 resini (13). Thrombin activity was measured in a fibrinogen-clotting assay (13) standardized with U. S. Standard Thrombin, Lot J, ol)tained from Dr. D. L. Aronson, Bureau of Biologics, Food and( Drug Administration, Bethesda, Md. Proteini coll- cenltratioin was determined by ultraviolet absorbance, as- sUminlg E28,'1' = 18.3 for thrombin (13). The specific activity Mr 32,000-'-W * ranlgedl fromii 2,400 to 2,700 U/mg thrombin. The preparationi was homogeneous by SDS polyacrylamide gel electrophoresis, Tracking dyee_ which indicated the absence of detectable quantities of 8- and y-thrombin (13). Humani AT III was purified by the method of Thaler anid Schmer (14). In a standard clotting assay, the prepara- FIGURE 1 Interaction of 1251-thrombin with and tion inhibited 1,500-2,000 U of thrombin/mg of AT III, plasma puri- the assumiption of = 5.7 for AT III The fied AT III. Samples contained either plasmiia (1 ul) or AT onl E280'% (15). AT III (1 ,ul of a 200-,ug/ml stock solution) in a total volume of III was homogeneous and had an Mr of 59,000 by SDS poly- 50 ,ul of 0.15 M NaCl, 0.05 M Tris-HCl, pH 7.4. 1251-thromiibin acrylamiidle gel electrophoresis. and heparin at final concentrations of 1 U/ml and 5 U/ml, Ioditnation of thrombitn. Thrombin was iodinated by the respectively, were also present as indicated. After a 10-mim chloramiiine-T method described previously (16), modified incubation at 25°C, each sample was prepared for electro- to include 20 ,uM unlabeled Nal in the reaction mixture. phoresis and autoradiography as described in Methods. The This resulted in the incorporation of -0.5 atom of I/molecule autoradiogram is shown. (a) '251-thrombin, (b) 1251-throllmbin of thrombin, and in a specific radioactivity of 2-4 x 106 cpm/U + plasma, (c) 125I-thrombin + plasmiia + heparin, (d) 1251_ thrombin. The 1251-thrombin had >90% of the specific clotting thrombin + purified AT III + heparin. activity of the unlabeled starting material. In parallel incubations conitaining 10 ,tg/ml of purified AT III and 0.08 ,g/ml of either '251-thrombin or unlabeled thrombin, the rates of in- activatioin of the two thrombins were identical (data not resulting inhibitor-protease complexes by SDS poly- shown). The second-order rate constant calculated for the acrylamide gel electrophoresis. In this manner we inactivation of 1251-thrombin by AT III was 5.3 x 105 M-1 have demonstrated the presence of a previously min-', in agreement with the value for unlabeled thrombin re- thrombin inhibitor with cofactor ported by others (3). unrecognized heparin SDS gel electrophoresis. Samples were heated at 100°C activity. The new inhibitor is immunologically dis- for 2 min with an equal volume of a solution containing tinct from AT III and from all of the other known 4% SDS, 10% ,f-mercaptoethanol, 20% glycerol, and 0.125 M protease inhibitors in plasma. The inhibitor is present Tris-HCI, pH 6.8. Electrophoresis was performed with slab in plasma at a concentration somewhat lower than gels containing 7.5% polyacrylamide in the running gel AT and 4% polyacrylamide in the stacking gel according to that of III. In the presence of relatively high the procedure of Laemmli (17). The gels were fixed and concentrations of heparin, however, thrombin forms stained with a solution containing 0.25% Coomassie Blue complexes preferentially with the new inhibitor. G-250, 10% acetic acid, and 50% methanol. The gels were destained with a solution containing 10% acetic acid and 25% ethanol, and dried with a Hoefer model SE540 slab gel METHODS dryer (Hoefer Scientific Instruments, San Francisco, Calif.). Materials. Blood was drawn from normal volunteers into Molecular weight standards included myosin (Mr 200,000) plastic syringes containing 6 mM EDTA (final concentra- ,8-galactosidase (Mr 116,500), phosphorylase B (M, 94,000), tion). Cells were removed by centrifugation, and the plasma bovine serum albumin (Mr 68,000), and ovalbumin (Mr 43,000). was stored at -20°C before use. Heparin derived from Autoradiography. Kodak RP X-Omat film (Eastman Kodak porcine intestinal mucosa (Panheprin, Abbott Diagnostics, Co., Rochester, N. Y.) was placed between the dried slab Diagnostic Products, North Chicago, Ill., 1,000 U. S. Phar- gel and a Dupont Cronex Lightning Plus intensifier screen macopeia U/ml; or Lipo-Hepin, Riker Laboratories, Inc., (E. I. Du Pont de Nemours & Co., Wilmington, Del.) Northridge, Calif., 20,000 U. S. Pharmacopeia U/ml) was for 6-18 h at -70°C (18), and the film was developed diluted in buffer containing 0.15 M NaCl and 0.05 M Tris- according to the manufacturer. Selected bands identified HCl, pH 7.4, immediately before use; the biological activ- in the autoradiogram were traced onto the dried gel, cut ity was assumed to be that stated by the manufacturer. out, and counted in a Beckman model 300 gamma counter According to the carbazole assay for uronic acid (10), the (Beckman Instruments, Inc., Fullerton, Calif.). specific activities of the two heparin preparations were Immunoabsorption of plasma. Heat-killed, formalin-fixed 159 U/mg (Abbott Diagnostics) and 144 U/mg (Riker). Rabbit Staphylococcus aureus (Cowan I strain) was provided by Dr. antisera directed against human AT III, al-antitrypsin, al- Benjamin Schwartz, Washington University, St. Louis, Mo.

590 D. M. Tollefsen and M. K. Blank Immediately before use, the bacteria were washed with 0.25% Nonidet P-40 in 0.15 M NaCl, 0.05 M Tris-HCl, pH 7.4, and resuspended in the same buffer without detergent at a concentration of 8-10% (packed cell volume). Bacteria were coated with antibodies specific for one of the plasma protease inhibitors by incubating 0.5 ml of the S. aureus sus- Mr 96,000 -- - pension with 0.5 ml of rabbit antiserum for 2 h at 25°C. The bacteria were washed in the above buffer (without de- Mr 85,000 a- tergent) to remove unbound serum proteins and then sedimented at 12,000 g for 10 min to produce a tightly packed pellet. 50 ,ul of human plasma was then mixed with the bacterial pellet (-50 ,ul, packed cell volume) and allowed to incubate for 2 h at 25°C. The bacteria were removed by centrifugation and the absorbed plasma was stored at -20°C. Protein concentrations of the absorbed plasma samples were measured by the Lowry method (19) and found to be -85% of that of the starting plasma, which indicates a minor degree of dilution of the plasma during the absorption procedure. Mr 32,000 - Ouchterlony analysis. Double immunodiffusion was performed with 1%-agarose plates prepared by standard meth- a b c d e f ods (20). Diffusion was allowed to proceed for 24-48 h in a FIGURE 2 Reaction of 1251-thrombin-inhibitor complexes with humidified chamber at 25°C, and the plates were then photo- antiserum to AT III. All incubations were performed at 250C graphed. In some cases, the plates were washed, dried, in buffer containing 0.15 M NaCl and 0.05 M Tris-HCI, pH and stained for protein with Coomassie Blue R-250. 7.4. Plasma (5 ,l) or purified AT III (5 Al of a 200-,ug/ml Immunoassay of AT III. AT III concentration in plasma solution) was incubated with 5 U/ml heparin and 1 U/ml was determined by radial immunodiffusion with agar plates 1251-thrombin in a total volume of 250 A.l. The 1251-thrombin containing antisera to human AT III (M-Partigen plates, was added last, after a 1-min preincubation of the other Calbiochem-Behring Corp.) (21). The assay was standardized reactants. After a 10-min incubation, 5 ilI of antiserum was with purified AT III. added. 30 min later, 50 Al of an 8% (vol/vol) suspension of staphylococci prepared as described in Methods was RESULTS added, and the tubes were stirred gently for an additional 30 min. The bacteria were removed by centrifugation 125I-Thrombin preparation. Thrombin consists of through 300 ,ul of 10% Apiezon A oil in nl-butyl phthalate two polypeptide chains linked by a disulfide bond. in an Eppendorf microcentrifuge, and the pellet was re- They include the A-chain (Mr 4,600) and the B-chain suspended in 300 ,uLl of buffer. Aliquots (50 ,l) of the 32,000), the latter containing the active center supernate (a, c, and e) and the resuspended pellet (b, d, and (M, f) were prepared for electrophoresis as described in Meth- serine residue of the protease (13). In reduced SDS ods. The autoradiogram is shown. (a and b) '251-thrombin polyacrylamide gels, the 1251-thrombin used in the ex- + plasma + anti-AT III antiserum, (c and d) 1251-thrombin periments to follow gave the pattern shown in Fig. + AT III + anti-AT III antiserum, (e and f) 1251-thrombin la. Most of the radioactivity (73-84% in different + plasma + anti-al-antitrypsin antiserum. preparations) was present as a single band at the position of the B-chain. The remaining 16-27% of the A different pattern was observed when 5 U/ml radioactivity migrated with the tracking dye and was heparin was included in the incubation (Fig. 1c). Under assumed to represent the dissociated A-chain. these conditions, 50% of the radioactivity was found Heparin-dependent complexes of '251-thrombin and in three complexes of Mr 72,000, 85,000, and 96,000. plasma proteins. When 1251-thrombin was incubated An additional 15% of the 1251 was distributed uni- with plasma (diluted 1:50) in the absence of heparin, formly throughout the area of the gel between the Mr- 20% of the 125I became incorporated into several com- 72,000 and Mr-32,000 bands (discussed below). There plexes of Mr 72,000-200,000 in reduced SDS gels was a concomitant reduction in the amount of radio- (Fig. lb). Bands representing the uncomplexed A- activity at the position of the thrombin B-chain (now and B-chains contained 17 and 63% of the 1251, only 18% of the total counts per minute), whereas the respectively. The complexes observed in this experi- radioactivity at the tracking dye did not change ment were not characterized further. 7% of the 1251, significantly when compared with the incubation with- however, was found in bands of Mr 72,000 and 85,000, out heparin. Control incubations containing 1251-throm- which probably represent complexes of 1251-thrombin bin and heparin without plasma did not form com- with AT III, as shown below. This is precisely the plexes (not shown). The observed complexes, there- extent of complex formation with AT III expected fore, appear to result from covalent bond formation after a 10-min incubation, on the assumption of a between plasma proteins and the B-chain of thrombin. second-order rate constant of 2.07 x 103 M-1 s-1 (22) The Mr-72,000 and Mr-85,000 bands in Fig. Ic con- and a concentration of 4,ug AT III/ml in the incubation tained 2/3 of the complexed B-chain radioactivity, mixture. and they comigrated with the products formed when

Heparin-dependent Protease Inhibitor 591 c

FIGURE 3 Ouchterlony analysis of inhibitor-depleted plassmas. Plasma samples were absorbed with antisera to specific protease inhibitors and immunodiffusion was performed as described in Methods. In each case, the center well contained 5 ,ul of the antiserum used to absorb the plasma; wells 1, 2, and 3 contained 5 gl of control unabsorbed plasma (undiluted, 1:2, and 1:4, respectively); and wells 4, 5, and 6 contained 5 ,lI of absorbed plasma (undiluted, 1:2, and 1:4). Antisera to the following inhibitors were used: a, a2-antiplasmin; b, al-antitrypsin; c, a2-macroglobulin; d, antithrombin III; e, Cl inactivator; f, al-antichymotrypsin; and g, inter-a-trypsin inhibitor. Plates b andf were dried and stained before being photographed.

1251-thrombin was incubated with an amount of puri- complex was quantitatively removed from plasma with fied AT III equal to that present in the plasma sample antiserum to AT III and protein A-bearing staphylococci (Fig. ld). The two bands had approximately the same (25), >90% of the Mr-96,000 complex remained in the apparent molecular weights as the intact thrombin-AT supernate (Fig. 2, a and b). Doubling the quantity III complex and the major degradation product of this of antiserum and staphylococci did not result in re- complex reported by others (23, 24). In subsequent moval of more of the M,-96,000 complex (not shown). experiments (Figs. 2, 4, and 5) heparin was preincu- The Mr-85,000 complex prepared from i25I-thrombin bated with plasma or AT III for 1 min before the addition of 125I-thrombin, and the Mr-72,000 band was no longer seen. The M,-96,000 band shown in Fig. lc contained -1/3 of the complexed B-chains in the presence of 5 U/ml heparin. The band was virtually undetectable in incubations without heparin (Fig. lb) and did not Mi r ii} form in the reaction of purified AT III and '25I- thrombin (Fig. ld). Since electrophoresis was car- ried out in the presence of SDS and ,8-mercaptoethanol, it is unlikely that the Mr-96,000 band is a degradation product of the Mr-85,000 thrombin-AT III complex. These results suggest that plasma contains in addition to AT III a second protein capable of forming a complex with thrombin in the presence of heparin. frcickrinq v e -- Lack of immunoreactive AT III in the Mr-96,000 complex. Alternatively, we considered that the Mr, 96,000 band might contain a precursor of AT III or FIGURE 4 Formation of 125I-thrombin-inhibitor complexes in result from interaction of the thrombin-AT III com- inhibitor-depleted plasmas. Plasma samples were incubated plex with a third plasma component. In either case with 10 U/ml 1251-thrombin and 5 U/ml heparin, and electro- the complex might have been found (although not phoresis was performed as in Fig. 1. The '251-thrombin was added after a 1-min preincubation ofthe other reactants. Lanes necessarily) to contain antigenic determinants in a -g of the autoradiogram correspond to the absorbed plasma common with native AT III. Fig. 2 demonstrates samples in plates a-g, respectively, in Fig. 3. Lane h repre- that this was not the case. Although the Mr-85,000 sents a control incubation with unabsorbed plasma.

592 D. M. Tollefsen and M. K. Blank and purified AT III was absorbed as expected under 40~ A identical conditions (Fig. 2, c and d). When nonim- 196K mune rabbit serum or antiserum against human al- 30k I 85K antitrypsin (Fig. 2, e and f) was substituted for anti- AT III antiserum, neither complex was found in the 10 132K bacterial pellet. - - ITD Formation of the Mr-96,000 complex in plasma c2i u C N D samples depleted of known protease inhibitors. We 196K next absorbed aliquots of plasma with antibodies I 85K directed against individual protease inhibitors (for de- cs 20f- tails see Methods), and we assayed each absorbed ]32K plasma for the ability to form complexes with 125I1 ITD thrombin. The effectiveness of the absorption procedure was first confirmed by double immunodif- 40 E 196K fusion analysis. Fig. 3 demonstrates that a 1:4 dilution 85K of unabsorbed plasma formed a visible precipitin line 30 - when allowed to react with antiserum to each of the 2C 32K inhibitors tested. In contrast, undiluted samples of - plasma that had been absorbed with antibodies against ITD -I -L-L N LI O Cr4 4 000 1 o o O o O - 0000E;6o- om o~o - ')C 00 - a2-antiplasmin, a2-macroglobulin, antithrombin III, o Q) o OO O o o Cl inactivator, al-antichymotrypsin, and inter-a-trypsin 0 0 inhibitor failed to form precipitin lines with the ab- HEPARIN CONCENTRATION (U/ml) sorbing antisera. When the - 15% dilution of the plasma FIGURE 5 Formation of 1251-thrombin-inhibitor complexes at samples during absorption (see Methods) is taken into various heparin concentrations. Unabsorbed plasma (A and account, Fig. 3 indicates substantial (i.e., >%) deple- B), purified AT III (C and D), and AT III-depleted plasma the inhibitors from these plasma samples. a-l- (E and F) were incubated with 1 U/ml 1251-thrombin as in tion of Fig. 1 after a 1-min preincubation with 0-100 U/ml heparin. antitrypsin, which is present in plasma at an 8- to 56- Electrophoresis and autoradiography were performed as fold greater molar concentration than any of the other described in Methods. Selected bands identified by auto- inhibitors (4), was not effectively removed by this radiography (bracketed) were cut out of the dried gels and technique (Fig. 3b). counted for 125I. The total 125j-thrombin B-chain counts per minute was assumed to be the sum of the counts per minute When the absorbed plasma samples were incubated found in bands above the tracking dye (TD). Panels A, with 125I-thrombin and heparin, the Mr-96,000 com- C, and E show the percentage of the total B-chain counts plex was formed in each case (Fig. 4, a-g). The Mr- per minute found in the following bands: 0, Mr-96,000 85,000 complex was also present as expected, ex- complex; *, Mr-85,000 complex; O, area between the M,- 85,000 and Mr-32,000 bands; 0, area between the Mr-96,000 cept in the plasma sample depleted of AT III (Fig. 4d). and Mr-32,000 bands. Panels B, D, and F show the auto- The radioactivity in the Mr-96,000 band in each of the radiograms. absorbed plasma samples was 71-93% of that present in unabsorbed plasma (Fig. 4h). It is important to note that a saturating concentration of 125I-thrombin concentrations, a faint Mr-96,000 band was also ob- (see below) was used to maximize complex formation served. Increasing the concentration of heparin from in these incubations. With the exception of al-anti- 0.1 to 20 U/ml resulted in a substantial increase in trypsin, for which the data are inconclusive, the ex- the amount of radioactivity in the Mr-96,000 band and periments in Figs. 3 and 4 indicate that the Mr-96,000 a concomitant decrease in the intensity of the Mr, complex does not result from interaction of 125I1 85,000 band. At concentrations >20 U/ml, the propor- thrombin with any of the known protease inhibitors tion of "-5I counts per minute in the two bands did of plasma. not change. Identical results were obtained with Heparin-dependence of the Mr-96,000 complex com- heparin preparations from two commercial sources pared with that of the thrombin-AT III complex. We (see Methods). determined the extent of complex formation when 1251_ Preliminary experiments in which 125I-thrombin thrombin was added to plasma, purified AT III, or AT (1 U/ml) was added to whole blood collected directly III-depleted plasma in the presence of various con- into syringes containing various amounts of concentrations of heparin. When a limiting amount of centrated heparin have yielded qualitatively similar '25I-thrombin was added to plasma (Fig. 5, a and b), results,3 although the concentrations of heparin re- the amount of radioactivity present in the Mr-85,000 thrombin-AT III complex was maximal at 0.01-0.1 U/ml of heparin. At these relatively low heparin 3D. M. Tollefsen and M. K. Blank, unpublished observation.

Heparin-dependent Protease Inhibitor 593 quired to activate both the new inhibitor and AT III assaying the same plasma samples by radial immuno- appeared to be two- to threefold greater than those diffusion. in Fig. 5, a and b. Studies in which purified AT III was substituted DISCUSSION for plasma (Fig. 5, c and d) indicated (a) that the extent of formation of the Mr-85,000 complex was maximal Our results indicate that human plasma contains a at 0.01 U/ml of heparin and did not decrease with substantial quantity of an unidentified heparin-de- higher heparin concentrations, and (b) that the Mr- pendent inhibitor of thrombin. Like antithrombin III 96,000 complex did not form even at the highest (23), the new inhibitor forms a complex with the concentrations of heparin. Additional experiments us- thrombin B-chain that is stable in the presence of ing AT III-depleted plasma under identical incubation SDS and P-mercaptoethanol at 100°C. We have shown conditions (Fig. 5, e and f) demonstrated maximal that thrombin treated with diisopropylfluorophosphate formation of the Mr-96,000 complex at a heparin con- (16) or hirudin (27) to block the proteolytic active centration of 1 U/ml; as in Fig. 4d, no band correspond- site of the molecule fails to form a complex with the ing to the thrombin-AT III complex was seen. new inhibitor in plasma.3 These experiments are Although not readily apparent from the autoradio- consistent with the formation of a covalent bond be- grams in Fig. 5, a significant amount of radioactivity tween the protease and the inhibitor and suggest was present in the area of the gels between the Mr- that the active site of thrombin is required to form the 85,000 and Mr-32,000 bands. As shown in Fig. 5 a, c, complex (cf. 23 and 28). and e (open symbols) the radioactivity in this region The new inhibitor differs from AT III in the follow- paralleled that present in the Mr-85,000 complex but ing respects: (a) It combines with thrombin to form not in the Mr-96,000 complex, and it may thus repre- a complex of Mr 96,000 in reduced SDS polyacryl- sent dissociation of a fraction of the thrombin-AT amide gel electrophoresis (Fig. 1); this complex III complexes (40-50%) during electrophoresis. This migrates more slowly than the thrombin-AT III com- phenomenon has been described previously and may plex (Mr 85,000). (b) The Mr-96,000 complex does be connected with the presence of nucleophiles in not form when thrombin is allowed to react with the Laemmli running gel (26). Initial attempts to purified AT III in the presence or absence of heparin. separate the two complexes by electrophoresis in (c) Antibodies directed against human AT III do not neutral phosphate buffers have been unsuccessful.3 react with the Mr-96,000 complex (Fig. 2). (d) Finally, Estimation of the quantity of the new heparin- the Mr-96,000 complex is formed when thrombin is dependent inhibitor in plasma. By increasing the incubated with plasma immunologically depleted of amount of 1251-thrombin incubated with a fixed quan- AT III (Figs. 4 and 5). tity of plasma in the presence of 5 U/ml heparin, Additional experiments established that the new we demonstrated saturation of both the Mr-85,000 and inhibitor is not identical to any of the other known plasma Mr-96,000 complexes in SDS gels. The amount of protease inhibitors. Thus, the Mr-96,000 complex forms thrombin incorporated into each complex was cal- in approximately normal amounts in plasma samples culated from the specific radioactivity of the 1251_ immunologically depleted of a2-antiplasmin, a2-macro- thrombin B-chain. Analysis of samples from 16 normal globulin, Cl inactivator, al-antichymotrypsin, or inter- subjects revealed the following: 46 + 13 ,ug (mean + SD) a-trypsin inhibitor (Fig. 4). Although we were unable of thrombin in the Mr-96,000 complex and 93+13 ,g to remove al-antitrypsin from plasma by this tech- of thrombin in the Mr-85,000 complex/ml of plasma nique, we have shown that the Mr-96,000 complex at saturation. The plasma concentrations of the new does not react with antibodies directed against al- inhibitor and AT III were then estimated from these antitrypsin (Fig. 2) and that the complex between values, on the assumption of a 1:1 molar complex thrombin and al-antitrypsin migrates slightly more of thrombin and inhibitor in each case and molecular rapidly in SDS gels than the thrombin-AT III complex.3 weights of 36,600 (13), 59,000 (see Methods), and In addition, the Mr-96,000 complex was found in the 72,000 (see Discussion) for thrombin, AT III, and the plasma of two patients with combined Factor V/VIII new inhibitor, respectively. Thus, we calculated a con- deficiency,3 which indicates that the complex is not centration of 150 +21 ,ug/ml for AT III and 90 +26 ,g/ml due to the inhibitor of protein C recently described for the second heparin cofactor. These values, par- by Marlar and Griffin (29). It is also clear that the ticularly that of AT III, should be considered mini- new inhibitor is not identical to "protease-nexin," mum estimates of the true plasma concentrations, the heparin-dependent thrombin-binding protein of as the dissociated '251-thrombin was ignored in this fibroblasts described by Baker et al. (30), since the calculation. Indeed, a 50% higher value for the con- apparent molecular weights of the two complexes centration of AT III (225±+32 ,tg/ml) was obtained by are quite different (96,000 vs. 68,000).

594 D. M. Tollefsen and M. K. Blank In the experiments illustrated in Fig. 5, we have Our results indicate that higher heparin concentra- allowed the reaction of a limiting amount of thrombin tions are required to activate the new inhibitor. with plasma to go to completion in the presence of Recently, Hoogendoorn et al. (32) have removed AT varying concentrations of heparin. The simplest inter- III from plasma by absorption to heparin-Sepharose pretation of the results is that at higher heparin and have found that the absorbed plasma still had a concentrations thrombin reacts more rapidly with the prolonged thrombin time and activated partial throm- new inhibitor than with AT III. Heterogeneous boplastin time in the presence of heparin. Unfor- heparin preparations derived from porcine intestinal tunately, nio attempt was made to quantify the capacity mucosa contain many molecules that are inactive of the absorbed plasma to inactivate thrombin. toward AT III (3). Whether the same subfractions of We recently have purified the protein that reacts heparin activate both AT III and the new inhibitor with thrombin to form the Alr-96,000 complex in remains to be determined. The observation that the plasma.4 The purified inhibitor has an Mr of 72,000 new inhibitor is fully activated at 1 U/ml heparini in in the Laemmli gel system and does not react with AT III-depleted plasma compared with 20 U/ml in antibodies to any of the recognized plasma protease unabsorbed plasma (Fig. 5, a and e) suggests inhibitors.2 It inhibits the proteolytic and amido- that the two antithrombins might be competing for lytic activity of thrombin by forming a 1:1 molar com- the same biologically active heparin species. Alterna- plex with the protease. The rate of complex forma- tively, as Jordan et al. (3) have pointed out, the rate tion increases >1,000-fold in the presence of heparin. at which AT III reacts with thrombin actuallv de- Stuidies are in progress to elucidate the specificity creases at high heparin concentrations. This might of the new inhibitor with respect to various proteases allow the reaction with the new inhibitor to predominate. and muciopolysaccharides. The physiological role of the At present, it is reasonable to assume that at levels nlew inhibitor remains to be defined. of heparin achieved therapeutically by continuous intravenous infusion (i.e., 0.1-1.0 U/ml) (1), AT III ACKNOWLEDGMENT is the major activated antithrombin in the plasma. The results in 5 support This research was supported by grants HLBI 14147 (Spe- shown Fig. this hypothesis. cialized Center for Research in Thrombosis) and HL 16634 It is unknown whether heparin secreted from mast from the National Institutes of Health. cells or cell-surface mucopolysaccharides activate the new inhibitor. Thus, the two heparin-dependent anti- REFERENCES thrombins may be activated under different circum- stances in vivo. 1. O'Reilly, R. A. 1980. Anticoagulant, antithrombotic, and thrombolytic drugs. In The Pharmacological Basis of Previous investigators have separated plasma pro- Therapeutics. A. G. Gilman, L. S. Goodman, and A. teins by ion-exchange chromatography or electro- Gilman, editors. Macmillan Publishing Co., Inc., New phoresis and have claimed to resolve heparin co- York. 6th edition. 1347-1366. factor activity from AT III (7-9). Although in general 2. Rosenberg, R. D. 1977. Biologic actions of heparin. in Semin. Hematol. 14: 427-439. agreement with these reports, our studies differ two 3. Jordan, R., D. Beeler, and R. Rosenberg. 1979. Fractiona- important respects. We added thrombin directly to tion of low molecular weight heparin species and their plasma and then separated the resulting complexes interaction with antithrombin. J. Biol. Chem. 254: 2902- under denaturing conditions, thus minimizing the 2913. likelihood of artifacts such as proteolysis that might 4. Laskowski, NI., Jr., and I. Kato. 1980. Protein inhibitors of proteinases. Attnu. Rev. Biochem. 49: 593-626. occur during the fractionation of plasma. Further- 5. Rent, R., R. Myhrman, B. A. Fiedel, and H. Gewurz. 1976. more, we characterized the complexes immunologically. Potentiation of Cl-esterase inhibitor activity by heparin. Our results are at variance with more recent re- Cliii. Exp. Iimtlunol. 23: 264-271. ports claiming that AT III accounts for virtually all of 6. Pensky, J., L. R. Levy, and I. H. Lepow. 1961. Partial purification of a serum inhibitor of Ci-esterase. J. Biol. the heparin-dependent antithrombin activity of plasma. Cheim. 236: 1674-1679. Rosenberg and Damus (23) reported that 92% of the 7. Ganrot, P. 0. 1969. Electrophoretic separation of two heparin cofactor activity of heat-defibrinated plasma thrombin inhibitors in plasma and serum. Scand. J. Clin. was removed by anti-AT III immunoglobulins. Pre- Lab. Invest. 24: 11-14. liminary experiments from our laboratory indicate 8. Briginshaw, G. F., and J. N. Shanberge. 1974. Iden- tification of two distinct heparin cofactors in huimian that the new heparin-dependent inhibitor is sub- plasmiia: separation and partial purification. Arch. Bio- stantially inactivated under the conditions of heat chem. Biophyxs. 161: 683-690. defibrination.3 Holmer et al. (31) have also used antibodies to remove AT III from plasma. They found es- sentially no heparin cofactor activity in the AT III- 'D. M. Tollefsen, D. W. Majerus, and M. K. Blank. Manu- depleted plasma when tested with 0.04 U/ml heparin. script in preparation.

Heparin-dependent Protease Inhibitor 595 9. Briginshaw, G. F., and J. N. Shanberge. 1974. Identifica- 21. Mancini, G., A. 0. Carbonara, and J. F. Heremans. tion of two distinct heparin cofactors in human plasma. 1965. Immunochemical quantitation of antigens by single II. Inhibition ofthrombin and activated Factor X. Thromb. radial immunodiffusion. Immunochemistry. 2: 235-254. Res. 4: 463-477. 22. Lau, H. K., and R. D. Rosenberg. 1980. The isolation 10. Bitter, T., and H. M. Muir. 1962. A modified uronic and characterization of a specific antibody population acid carbazole reaction. Anal. Biochem. 4: 330-334. directed against the thrombin-antithrombin complex. 11. Miletich, J. P., G. J. Broze, Jr., and P. W. Majerus. J. Biol. Chem. 255: 5885-5893. 1980. The synthesis of sulfated dextran beads for isolation 23. Rosenberg, R. D., and P. S. Damus. 1973. The purifica- of human plasma coagulation Factors II, IX, and X. tion and mechanism of action of human antithrombin- Anal. Biochem. 105: 304-310. heparin cofactor. J. Biol. Chem. 248: 6490-6505. 12. Owen, W. G., and C. M. Jackson. 1973. Activation of 24. Fish, W. W., K. Orre, and I. Bjork. 1979. Routes of prothrombin with Oxyuranus scutellatus scutellatus thrombin action in the production of proteolytically (Taipan snake) venom. Thromb. Res. 3: 705-714. modified, secondary forms of antithrombin-thrombin 13. Fenton, J. W., II, M. J. Fasco, A. B. Stackrow, D. L. complex. Eur. J. Biochem. 101: 39-44. Aronson, A. M. Young, and J. S. Finlayson. 1977. 25. Kessler, S. W. 1976. Cell membrane antigen isolation Human thrombins: production, evaluation, and properties with the staphylococcal protein A-antibody adsorbent. of a-thrombin. J. Biol. Chem. 252: 3587-3598. J. Immunol. 117: 1482-1490. 14. Thaler, E., and G. Schmer. 1975. A simple two-step 26. Jesty, J. 1979. Dissociation of complexes and their deriva- isolation procedure for human and bovine antithrombin tives formed during inhibition of bovine thrombin and II/III (heparin cofactor): a comparison of two methods. activated Factor X by antithrombin III.J. Biol. Chem. 254: Br. J. Haematol. 31: 233-243. 1044-1049. 15. Kurachi, K., G. Schmer, M. A. Hermodson, D. C. Teller, 27. Bagdy, D., E. Barabas, L. Graf, T. E. Petersen, and S. and E. W. Davie. 1976. Characterization of human, Magnusson. 1976. Hirudin. Methods Enzymol. 45: 669- bovine, and horse antithrombin III. Biochemistry. 15: 678. 368-373. 28. Owen, W. G. 1975. Evidence for the formation of an ester between thrombin and heparin cofactor. Biochim. 16. Tollefsen, D. M., J. R. Feagler, and P. W. Majerus. 1974. Biophys. Acta. 405: 380-387. The binding of thrombin to the surface of human 29. Marlar, R. A., and J. H. Griffin. 1980. Deficiency of platelets. J. Biol. Chem. 249: 2646-2651. protein C inhibitor in combined Factor V/VIII deficiency 17. Laemmli, U. K. 1970. Cleavage of structural proteins disease. J. Clin. Invest. 66: 1186-1189. during the assembly of the head of bacteriophage T4. 30. Baker, J. B., D. A. Low, R. L. Simmer, and D. D. Nature (Lond.). 227: 680-685. Cunningham. 1980. Protease-nexin: a cellular component 18. Swanstrom, R., and P. R. Shank. 1978. X-ray intensifying that links thrombin and plasminogen activator and screens greatly enhance the detection by autoradiography mediates their binding to cells. Cell. 21: 37-45. of the radioactive isotopes 32P and 1251. Anal. Biochem. 31. Holmer, E., G. Soderstrom, and L-O. Andersson. 1980. 86: 184-192. Properties of antithrombin III depleted plasma. I. Ef- 19. Lowry, 0. H., N. J. Rosebrough, A. L. Farr, and R. J. fect of heparin. Thromb. Res. 17: 113-124. Randall. 1951. Protein measurement with the Folin 32. Hoogendoorn, H., A. Cerskus, F. Ofosu, M. Blajchman, phenol reagent. J. Biol. Chem. 193: 265-275. and J. Hirsh. 1980. Preparation and partial characteriza- 20. Garvey, J. S., N. E. Cremer, and D. H. Sussdorf. 1977. tion of human plasma depleted of antithrombin III by Methods in Immunology. W. A. Benjamin, Inc., Reading, heparin-Sepharose affinity chromatography. Thromb. Res. Mass. 3rd edition. 313-321. 20: 77-83.

596 D. M. Tollefsen and M. K. Blank